stack/tests/llm/test_config.py

"""llm.config — [llm] section parsing + env overrides."""

import pytest

from llm import config as llm_config


class TestLoad:
    def test_defaults_from_stack_toml(self, monkeypatch):
        monkeypatch.delenv("LLM_OLLAMA_HOSTS", raising=False)
        monkeypatch.delenv("LLM_PG_HOST", raising=False)
        monkeypatch.delenv("LLM_BUILD_ANN_INDEX", raising=False)
        cfg = llm_config.load()
        assert cfg.ollama_hosts == ("http://127.0.0.1:11434",)
        assert cfg.embed_model == "nomic-embed-text"
        assert cfg.embed_dim == 768
        assert cfg.pg_host == "127.0.0.1"
        assert cfg.pg_db == "llm"
        assert cfg.build_ann_index is True

    def test_env_overrides(self, monkeypatch):
        monkeypatch.setenv(
            "LLM_OLLAMA_HOSTS",
            "http://rig:11434, http://laptop:11434",
        )
        monkeypatch.setenv("LLM_PG_HOST", "postgres")
        cfg = llm_config.load()
        assert cfg.ollama_hosts == ("http://rig:11434", "http://laptop:11434")
        assert cfg.pg_host == "postgres"

    def test_build_ann_index_env_override_true(self, monkeypatch):
        monkeypatch.setenv("LLM_BUILD_ANN_INDEX", "1")
        cfg = llm_config.load()
        assert cfg.build_ann_index is True

    def test_build_ann_index_env_override_false(self, monkeypatch):
        monkeypatch.setenv("LLM_BUILD_ANN_INDEX", "no")
        cfg = llm_config.load()
        assert cfg.build_ann_index is False


class TestPgUrl:
    def test_url_includes_password_from_env(self, monkeypatch):
        monkeypatch.setenv("LLM_DB_PASSWORD", "s3cret")
        cfg = llm_config.load()
        url = llm_config.pg_url(cfg)
        assert url == f"postgresql+psycopg://llm:s3cret@{cfg.pg_host}:5432/llm"

    def test_missing_password_raises(self, monkeypatch):
        monkeypatch.delenv("LLM_DB_PASSWORD", raising=False)
        cfg = llm_config.load()
        with pytest.raises(RuntimeError, match="LLM_DB_PASSWORD"):
            llm_config.pg_url(cfg)


class TestParseHosts:
    def test_bare_and_annotated(self):
        out = llm_config.parse_hosts(
            "http://ollama:11434, http://rig.local:11434@24 ,http://nb:11434@12"
        )
        assert out == (
            ("http://ollama:11434", 0.0),
            ("http://rig.local:11434", 24.0),
            ("http://nb:11434", 12.0),
        )

    def test_trailing_slash_stripped_and_empty_parts_dropped(self):
        assert llm_config.parse_hosts("http://a:1/@8,,") == (("http://a:1", 8.0),)


class TestNewKnobs:
    def test_defaults_when_section_lacks_keys(self, monkeypatch):
        monkeypatch.delenv("LLM_OLLAMA_HOSTS", raising=False)
        cfg = llm_config.load()
        assert cfg.large_min_vram_gb == 20.0
        assert cfg.chat_num_ctx == 8192
        assert cfg.chat_chars_per_token == 2.4
        assert cfg.recency_half_life_days == 365.0
        assert cfg.recency_weight == 0.3
        assert cfg.k_per_kind == {"comment": 8, "rule": 4, "corpus": 4}
        assert cfg.top_n == 8
        assert isinstance(cfg.instruct_model_large, str)

    def test_env_hosts_populate_vram_map(self, monkeypatch):
        monkeypatch.setenv(
            "LLM_OLLAMA_HOSTS", "http://rig:11434@24,http://laptop:11434@12"
        )
        cfg = llm_config.load()
        assert cfg.ollama_hosts == ("http://rig:11434", "http://laptop:11434")
        assert cfg.host_vram == {"http://rig:11434": 24.0, "http://laptop:11434": 12.0}

    def test_valuation_knobs_defaults(self, monkeypatch):
        monkeypatch.delenv("LLM_DUCKDB_REPLICA", raising=False)
        cfg = llm_config.load()
        assert cfg.duckdb_replica == "data/replica/aco.ro.duckdb"
        assert cfg.valuation_years == 4
        assert cfg.code_cited_per_code == 3
        assert cfg.code_cited_collections == ("rules", "comments", "corpus")
        assert cfg.code_cited_max == 12
        assert cfg.lineage_max_rows == 25
        assert cfg.lineage_on_demand_max == 3

    def test_duckdb_replica_env_override(self, monkeypatch):
        monkeypatch.setenv("LLM_DUCKDB_REPLICA", "/app/data/replica/aco.ro.duckdb")
        assert llm_config.load().duckdb_replica == "/app/data/replica/aco.ro.duckdb"


class TestStrTuple:
    """``code_cited_collections`` must never explode a bare scalar string
    into a tuple of its characters (``tuple("rules")``)."""

    def test_scalar_string_becomes_one_tuple(self):
        assert llm_config._str_tuple("rules") == ("rules",)

    def test_list_passes_through_as_tuple(self):
        assert llm_config._str_tuple(["rules", "comments"]) == ("rules", "comments")

    def test_code_cited_collections_scalar_string_in_toml_is_one_tuple(
        self, monkeypatch
    ):
        import conf
        from conf import _Cfg

        raw = dict(conf.cfg.to_dict())
        raw["llm"] = dict(raw["llm"])
        raw["llm"]["code_cited_collections"] = "rules"
        monkeypatch.setattr(conf, "cfg", _Cfg(raw))
        assert llm_config.load().code_cited_collections == ("rules",)


def test_iom_companion_max_defaults_to_four():
    from llm.config import LlmConfig

    cfg = LlmConfig(
        ollama_hosts=("http://h",),
        embed_model="e",
        instruct_model="i",
        embed_dim=1,
        build_ann_index=False,
        pg_host="h",
        pg_port=1,
        pg_db="d",
        pg_user="u",
    )
    assert cfg.iom_companion_max == 4


def test_instruct_model_env_override(monkeypatch):
    # Judge-model comparisons (14b vs 32b) without editing stack.toml.
    from llm.config import load

    monkeypatch.setenv("LLM_INSTRUCT_MODEL", "qwen2.5:32b")
    assert load().instruct_model == "qwen2.5:32b"
    monkeypatch.delenv("LLM_INSTRUCT_MODEL")
    assert load().instruct_model == "qwen2.5:14b"


def test_tagging_parameters_load_from_config():
    # `stack llm tag` defaulted to min_confidence 0.0 while eval-tags used
    # 0.60: the run and its gate disagreed. Both now read these.
    from llm.config import load

    cfg = load()
    assert cfg.tag_top == 12
    assert cfg.tag_max_tags == 3
    assert cfg.tag_min_confidence == 0.45